SEARCH

SEARCH BY CITATION

Keywords:

  • epidermal growth factor receptor;
  • lung cancer;
  • adenocarcinoma;
  • surgical pathology;
  • cytopathology

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND

Activating mutations in the epidermal growth factor receptor (EGFR) in non–small cell lung carcinoma (NSCLC) are associated significantly with responsiveness to EGFR tyrosine kinase inhibitors. The objective of this study was to investigate the suitability of cytologic specimens for assessing EGFR mutations in lung adenocarcinomas.

METHODS

Sixty paired histologic and cytologic specimens of lung adenocarcinoma were collected. Exons 18 through 21 of the EGFR gene were amplified using polymerase chain reaction, and the mutation status of each sample was analyzed by pyrosequencing. A comparison of EGFR mutation status between histologic specimens and cytologic specimens was performed.

RESULTS

The overall EGFR mutation concordance rate between histologic specimens and corresponding cytologic specimens was 91.7%. No significant difference was observed in the concordance rate between cytologic specimens from primary lesions and specimens from metastatic lesions (P = .63). The following parameters were correlated with the most reliable EGFR mutation results using the pyrosequencing method (100% concordance with the corresponding histologic specimens) in cytologic samples: a DNA concentration >25 ng/μL, content of >30 tumor cells, or a tumor percentage >30%.

CONCLUSIONS

In this study, routinely prepared cytologic specimens were reliable sources for assessing EGFR mutation status. The authors concluded that cytologic specimens from metastatic lesions and primary tumors are suitable for the successful assessment of EGFR mutation status. Cancer (Cancer Cytopathol) 2013;121:311–9. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

Major discoveries in the molecular pathogenesis of lung cancer have resulted in the successful application of targeted therapeutic strategies.[1] One example of these successes is a subset of patients who had lung cancer with epidermal growth factor receptor (EGFR) mutations and exhibited improved clinical responsiveness to tyrosine kinase inhibitors.[4, 6] Thus, the molecular evaluation of EGFR mutations has gained increasing clinical importance in daily practice.

Most patients with non–small cell lung carcinoma (NSCLC) present with advanced-stage disease at the time of diagnosis, and surgical treatment is not recommended. In these patients, the diagnosis of lung cancer often is based on cytologic specimens, such as fine-needle aspiration (FNA), pleural fluid, and bronchial washing or brushing specimens, which may represent the only material available for EGFR mutation analysis. Therefore, determining whether cytology specimens are adequate for analyzing EGFR mutation status is of crucial clinical importance.

Several studies have examined the EGFR mutation status of tumors using cytologic specimens and have demonstrated a similar or higher frequency of mutations in these specimens than in surgical specimens.[7] However, few of those studies have compared histologic specimens with cytologic specimens from the same patient. Thus, it remains unclear whether the mutation status in cytologic specimens can reflect the profiles revealed by histologic specimens from the same patient. In addition, several studies have reported on the EGFR mutation status in primary and metastatic lesions. However, comparisons of the mutation status between primary tumors and metastatic lesions using cytologic specimens have not been reported.[11]

Various types of cytologic preparations have been used to assess EGFR mutations. A recent review indicated that the most commonly used type of cytologic specimen for assessing EGFR mutations is formalin-fixed, paraffin-embedded cell blocks followed by archival smear slides and fresh cells.[14] Because cell blocks and archival smear slides represent the most commonly used preparations, an evaluation of the suitability of these 2 specimens for EGFR mutation analysis is necessary. The objectives of the current study were to compare the assessment of EGFR mutation status between histologic and cytologic specimens from lung adenocarcinomas and to investigate whether these cytologic specimens are suitable for the molecular assessment of EGFR mutation status.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

Sample Collection

We retrospectively reviewed and assessed the EGFR mutation status of 382 surgically resected or biopsied primary NSCLC tumors at our university hospital between October 2006 and March 2012. We sought cases with available cytologic specimens, such as cytologic aspirates from the primary tumor, bronchial washing, sputum, and mediastinal lymph node or pleural fluid metastatic specimens. In total, 60 cases were evaluated.

The 60 histologic specimens consisted of either biopsy materials (11 transbronchial biopsy and 22 fine-needle biopsy specimens) or resection materials (27 cases). The 60 cytologic specimens included 36 primary lung carcinomas and 24 metastatic lesions. The 36 primary specimens included 31 FNA specimens, 2 bronchial washings, and 3 sputum cytologic specimens. The 24 metastatic lesions included 7 lymph node FNA specimens, 16 pleural effusions, and 1 ascites specimen. Of the 60 cytologic specimens, 43 were available as both cell blocks and archival smear slides for the comparison of EGFR mutation status. This study was approved by our institutional review board.

Cytologic Specimen Selection for EGFR Mutation Testing

All original slides were reviewed independently by 2 pathologists (P.L.-S. and J.-H.C.). The tumor percentage was evaluated based on the number of tumor cells relative to all nucleated cells within each specimen, and tumors were graded according to their conformity into 3 categories of tumor percentages: 30%, 31% to 50%, and >50%. The slides also were assessed for the number of tumor cells and were graded according to their conformity into 4 categories: 30 cells, 31 to 50 cells, 51 to 100 cells, and >100 cells. The cell blocks were evaluated using the same method.

DNA Extraction

In histologic specimens and cell block specimens, genomic DNA was extracted from formalin-fixed, paraffin-embedded tissue samples as described previously.[15, 16] After deparaffinization using xylene, the tissue sections were stained using hematoxylin and eosin, and the target lesions were selectively dissected to minimize any non-neoplastic cell contamination. The QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) was used according to the manufacturer's protocols to isolate genomic DNA. A prolonged (48-hour) proteinase K digestion was performed on all of the formalin-fixed, paraffin-embedded tissues; this long digestion time releases the amplifiable nucleic acids by reversing formalin-induced crosslinks.[17] For the archival slides, DNA was extracted as follows: The area containing the greatest proportion of tumor cells was marked on the underside of each slide using a marking pen. Next, the slides from each specimen were incubated in xylene overnight to remove the coverslips. After removing the coverslips from the archival slides, the tumor cells were removed by scraping and were extracted for DNA. Briefly, 20 to 50 μL DNA extraction buffer solution (50 mM Tris buffer, pH 8.3; 1 mM ethylene diamine tetracetic acid, pH 8.0; 5% Tween-20; and 100 mg/mL proteinase K) with 10% resin were added to the scraped cells, and the cells were incubated at 56°C for a minimum of 1 hour. Next, the tubes were heated to 100°C for 10 minutes and centrifuged to pellet the debris. Five microliters of the supernatant were used for each polymerase chain reaction (PCR). A representative case is illustrated in Figure 1.

image

Figure 1. Epidermal growth factor receptor (EGFR) mutation analysis procedures are illustrated. (a) Target cells are marked. (b) Slides are incubated in xylene overnight to remove the coverslips. (c) Atypical cells of interest are scraped with a needle under the microscope. (d) A direct smear sample of pleural fluid is shown. (e) DNA extraction and polymerase chain reaction (PCR) are illustrated. (f) Electrophoresis is performed to determine the presence of PCR products. (g) Pyrosequencing analysis reveals an exon 19 deletion within the EGFR gene. (h) Pyrosequencing analysis reveals an exon 21 point mutation within the EGFR gene.

Download figure to PowerPoint

Polymerase Chain Reaction Amplification and Pyrosequencing Mutation Analysis

EGFR mutations in exons 18 through 21 were examined using PCR and a pyrosequencing method as previously described.[18] Each PCR mix contained forward and reverse primers (each 20 pmoL), 20 mmoL each of dinucleotide triphosphate, 50 mmoL MgCl2, 10 times PCR buffer, 50 U/μL Immolase DNA polymerase (Bioline USA Inc., Taunton, Mass), and 5 μL genomic DNA in a total volume of 50 μL. The PCR products were resolved by 3% agarose gel electrophoresis to confirm successful amplification of the PCR product. Forty microliters of each PCR product were bound to streptavidin Sepharose HP (GE Healthcare, Uppsala, Sweden), purified, washed, denatured using a 0.2 mol/L NaOH solution, and washed again. Next, 0.3 μmol/L of the pyrosequencing primers was annealed to the purified single-stranded PCR products, and pyrosequencing was performed using a PyroMark ID system (Qiagen) according to the manufacturer's instructions.

Statistical Analysis

Statistical analyses, including Fisher exact tests and the chi-square tests, were performed using SPSS software (SPSS Statistics 18.0; SPSS Inc., Chicago, Ill). Correlations between DNA concentration, tumor cell numbers, and tumor percentage were analyzed using Pearson correlation. Statistical significance was designated using a 2-tailed P value < .05.

We evaluated the suitability of the cytologic specimens for assessing EGFR mutation status by calculating the sensitivity, specificity, and diagnostic accuracy. Results from the EGFR mutation assessment using histologic specimens were considered the gold standard in this study. To calculate true-positive (TP), true-negative (TN), false-positive (FP), and false-negative (FN) accuracy, we used the following formulas: sensitivity = TP/(TP + FN) × 100; specificity = TN/(TN + FP) × 100; and accuracy = (TP + TN)/(TP + TN + FP + FN) × 100, respectively.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

Evaluation of the Cytology Specimens

Cytologic specimens were assessed for the quality and quantity of tumor tissue available for molecular testing and were evaluated in terms of 3 parameters: DNA concentration, estimated total tumor cell numbers, and approximate tumor percentage (Table 1). The concentration of DNA extracted from the archival smear slides ranged from 5.4 ng/μL to 745.5 ng/μL (mean, 159.5 ng/μL; median, 89.4 ng/μL). The concentration of DNA extracted from the cell blocks ranged from 3.8 ng/μL to 324.6 ng/μL (mean, 59.6 ng/μL; median, 40.7 ng/μL). Evaluation of the tumor cells indicated that estimated total tumor cell numbers ranged from 15 to 1200 cells (mean, 411.9 cells; median, 300 cells) in archival smear slides and from 10 to 1000 cells (mean, 231.9 cells; median, 90.0 cells) in cell blocks. The percentage of tumor cells ranged from 20% to 85% (mean, 66.5%; median, 70%) in archival smear slides and from 10% to 75% (mean, 48.3%; median, 50%) in cell blocks. The archival smear slide specimens contained significantly higher DNA concentrations, cell numbers, and tumor percentages than the cell block specimens (P < .001). The DNA concentrations correlated with both tumor cell numbers (archival smear slides: correlation coefficient [R] = 0.873; P < .001; cell blocks: R = 0.879; P < .001) and tumor percentages (archival smear slides: R = 0.391; P = .002; cell blocks: R = .816; P < .001) in both smear slides and cell blocks (Fig. 2). Among the different types of cytologic specimens, the FNA specimens contained significantly higher DNA concentrations and cell numbers than did the fluid specimens (P = .019 and P = .001, respectively).

Table 1. Specimen Characteristics
  DNA Concentration, ng/μL Cell Numbers Tumor Percentage 
   No. (%)  No. (%)  No. (%) 
SpecimenTotal No.Mean [Median]≤2526-50>50PMean [Median]≤3031-5051-100>100PMean [Median]≤30%31%-50%>50%P
  1. Abbreviations: FNA, fine-needle aspiration.

  2. a

    P for smear slides vs cell blocks.

  3. b

    P for FNAs vs pleural fluids and others.

Smear slides60159.5 [89.4]10 (16.7)10 (16.7)40 (66.7)0a411.9 [300]7 (11.7)7 (11.7)5 (8.3)41 (68.3)0a66.5 [70]4 (6.7)6 (10)50 (83.3)0a
FNA38 3 (7.9)9 (23.7)26 (68.4).019b 0 (0)7 (18.4)3 (7.9)28 (73.7).001b 1 (2.6)5 (13.2)32 (84.2).167b
Pleural fluid16 5 (31.3)0 (0)11 (68.8)  5 (31.3)0 (0)1 (6.3)10 (62.5)  1 (6.3)1 (6.3)14 (87.5) 
Others6 2 (33.3)1 (16.7)3 (50)  2 (33.3)0 (0)1 (16.7)3 (50)  2 (33.3)0 (0)4 (66.7) 
Cell blocks4359.6 [40.7]13 (30.2)14 (32.6)16 (37.2) 231.9 [90]11 (25.6)6 (14)7 (16.3)19 (44.2) 48.3 [50]10 (23.3)12 (27.9)12(27.9) 
FNA30 5 (16.7)12 (40)13 (43.3).013b 5 (16.7)3 (10)6 (20)16 (53.3).075b 5 (16.7)7 (23.3)18 (60).077b
Pleural fluid11 6 (54.5)2 (18.2)3 (27.3)  6 (54.5)1 (9.1)1 (9.1)3 (27.3)  4 (36.4)4 (36.4)3 (27.3) 
Others2 2 (100)0 (0)0 (0)  0 (0)2 (100)0 (0)0 (0)  1 (50)1 (50)0 (0) 
image

Figure 2. Correlations between DNA concentration, tumor cell numbers, and tumor percentage in smear slides (n = 60) and cell blocks (n = 43) are illustrated. Graphs illustrate (a) the correlation between DNA concentration and tumor cell numbers in smear slides (correlation coefficient [R] = 0.873; P < .001), (b) the correlation between DNA concentration and tumor cell numbers in cell blocks (R = 0.879; P < .001), (c) the correlation between DNA concentration and tumor percentage in smear slides (R = 0.391; P = .002), and (d) the correlation between DNA concentration and tumor percentage in cell blocks (R = 0.816; P < .001). Pearson correlation and statistical significance was designated using 2-tailed P values < .05.

Download figure to PowerPoint

Concordance in EGFR Mutations Between Histologic and Corresponding Cytologic Specimens

The EGFR mutation test results from 60 histologic and corresponding cytologic specimens are provided in Table 2. The overall concordance rate in EGFR mutation status between histologic specimens and their corresponding cytologic specimens was 91.7% (55 of 60 specimens) (Table 3). The FNA specimens exhibited 100% concordance with the corresponding histologic specimens. EGFR mutation status was discordant in 11 cases despite repeated analysis (Table 4). The 11 discordant cases included 4 pleural effusions, 5 FNA specimens, and 2 sputum specimens. In the specimens that fulfilled 1 of the 3 criteria (ie, >25 ng/μL DNA concentration, >30 tumor cell content, or >30% tumor percentage), the cytologic specimens exhibited 100% concordance with the corresponding histologic specimens.

Table 2. EGFR Mutation Status in Paired Histologic and Cytologic Specimens
    DNA Concentration, ng/μLNo. of CellsTumor Percentage, %EGFR Mutation Status
Patient No.HS TypeProcedure TypeProcedure SiteSSCBSSCBSSCBHSSSCB
  1. Abbreviations: ADC, adenocarcinoma; BW, bronchial washing; CB, cell blocks; E, exon; EBUS-FNA, endobronchial ultrasound-guided fine-needle aspiration; EGFR, epidermal growth factor receptor; FNA, fine-needle aspiration; FNB, fine needle biopsy; HS, histologic specimens; LN, lymph node; Met, metastatic; NA, not available; NSCLC, non-small cell lung cancer; SM, smear slides; TBB, transbronchial biopsy; WT, wild type; the shaded cells, the discordant cases.

1FNBBWBronchus582.37.81000506025WTWTWT
2FNBBWBronchus32.710065WTWTNA
3FNBSputumSputum125.517.9700407050E19E19E19
4FNBSputumSputum18.42530E21WTNA
5ResectionSputumSputum13.92520E21WTNA
6FNBFNALung3520.560307520E21E21WT
7ResectionFNALung486.538.97002006055WTWTWT
8TBBFNALung89.656.35204007070WTWTWT
9FNBFNALung42.112.590206020E21E21WT
10ResectionFNALung44.88.9110207020E19E19WT
11FNBFNALung26.830.850705050E21E21E21
12FNBFNALung154.9104.88005008070E21E21E21
13ResectionFNALung5024.5150307030E21E21WT
14FNBFNALung44.5125.62603107560E21E21E21
15ResectionFNALung27.46.340157010E21E21WT
16ResectionFNALung105.254600506070E19E19E19
17FNBFNALung114.830.6600803560E19E19E19
18ResectionFNALung125.382.16007008040E19E19E19
19ResectionFNALung74.925.4290456050E19E19E19
20ResectionFNALung89.234.4350908060E19E19E19
21FNBFNALung21.8109.3601204055E19E19E19
22ResectionFNALung56543.910001507045E19E19E19
23TBBFNALung53.640.7185808050E19E19E19
24ResectionFNALung187.134.5700808060E19E19E19
25ResectionFNALung64189.94008007560E19E19E19
26FNBFNALung107117.56001508065E19E19E19
27FNBFNALung108.242.38001408560E19E19E19
28ResectionFNALung12045.64001504560E19E19E19
29ResectionFNALung68.119065E19E19NA
30ResectionFNALung114.225065E21E21NA
31TBBFNALung83.430070E21E21NA
32ResectionFNALung11130080E21E21NA
33ResectionFNALung31.84575E21E21NA
34ResectionFNALung124.555075E21E21NA
35TBBFNALung13.53570E21E19NA
36ResectionFNALung126.360070WTWTNA
37ResectionFluidAscites85.720080E19E19NA
38TBBFluidPleura5.93.830307045WTWTWT
39FNBFluidPleura5.45.925207040E21E21E21
40TBBFluidPleura8.217.615256020E19WTWT
41FNBFluidPleura12.812.220252530E19WTWT
42TBBFluidPleura745.5119.312006006060E21E21E21
43FNBFluidPleura261.75.1400308015E19E19E19
44ResectionFluidPleura123.133.1300507045E19E19E19
45FNBFluidPleura514.7100108010E21E21WT
46FNBFluidPleura110.338.9250608550E19E19E19
47TBBFluidPleura700.1201.1110010008070E19E19E19
48FNBFluidPleura453102.810007506055E19E19E19
49ResectionFluidPleura10.12050E19WTNA
50FNBFluidPleura558.5100070E21E21NA
51ResectionFluidPleura168.780080E21E21NA
52FNBFluidPleura622.710008019&2119&21NA
53ResectionFluidPleura123.180072WTWTNA
54ResectionFNALN26.962.1404007050WTWTWT
55TBBFNALN2073.6402007060E19E19E19
56FNBFNALN86.647.86001007060E19E19E19
57TBBEBUS-FNALN6942.52901507040E19E19E19
58ResectionEBUS-FNALN30855.65005004060E19E19E19
59TBBEBUS-FNALN266107.88007006075E19E19E19
60ResectionEBUS-FNALN564.8324.6100010006075E19E19E19
Table 3. EGFR Mutation Concordance in Different Specimens
Comparison Between Different SpecimensTotal No.Concordance: No. (%)P
  1. Abbreviations: EGFR, epidermal growth factor receptor gene; FNA, fine-needle aspiration.

Histologic vs cytologic specimens6055 (91.7) 
FNA3838 (100) 
Pleural fluid1613 (81.3) 
Bronchial washing22 (100) 
Sputum31 (33.3) 
Ascites11 (100) 
EGFR mutation in different tumor site  .63
Histologic vs cytologic specimens from primary tumor site3634 (94.4) 
FNA, lung3131 (100) 
Bronchial washing22 (100) 
Sputum31 (33.3) 
Histologic vs cytologic specimens from metastatic site2421 (87.5) 
FNA, lymph node77 (100) 
Pleural fluid1613 (81.3) 
Ascites11 (100) 
Histologic specimens vs 2 different cytology preparations   
Smear slides vs histologic specimens4341 (95.3).04
Cell blocks vs histologic specimens4335 (81.4) 
Table 4. Discordant Cases
     DNA Concentration, ng/μLNo. of Tumor CellsTumor Percentage, %EGFR Mutation Status
Patient No.HS TypeProcedure TypeProcedure SiteCytology DiagnosisSSCBSSCBSSCBHSSSCB
  1. Abbreviations: ADC, adenocarcinoma; CB, cell blocks; E, exon; EGFR, epidermal growth factor receptor; FNA, fine-needle aspiration; FNB, fine-needle biopsy; HS, histologic specimens; Met, metastatic; NA, not available; NSCLC, non-small cell lung cancer; SM, smear slides; TBB, transbronchial biopsy; WT, wild type; the dark and light shading cells, the discordant cases in smear slides and cell blocks, respectively.

4FNBSputumSputumNSCLC18.42530E21WTNA
5ResectionSputumSputumNSCLC13.92520E21WTNA
6FNBFNALungADC3520.560307520E21E21WT
9FNBFNALungADC42.112.590206020E21E21WT
10ResectionFNALungADC44.88.9110207020E19E19WT
13ResectionFNALungFavor ADC5024.5150307030E21E21WT
15ResectionFNALungFavor ADC27.46.340157010E21E21WT
40TBBFluidPleuraMet ADC8.217.615256020E19WTWT
41FNBFluidPleuraAtypical cell12.812.220252530E19WTWT
45FNBFluidPleuraMet ADC514.7100108010E21E21WT
49ResectionFluidPleuraMet ADC10.12050E19WTNA

The Concordance Rate of EGFR Mutations in Cytologic Specimens From Primary Tumor Tissues and Metastatic Lesions

To determine the reliability of the EGFR mutation assessment performed on cytologic specimens from metastatic lesions compared with specimens from primary tumor sites, we compared the concordance rate of EGFR mutations between the histologic specimens and the corresponding cytologic specimens from primary tumor sites with the rate of cytologic specimens from the metastatic lesions. The concordance rates between the histologic and corresponding cytologic specimens were 94.4% (34 of 36 specimens) and 87.5% (21 of 24 specimens) in primary lesions and metastatic lesions, respectively. No significant difference was observed in the concordance rate between cytologic specimens from primary lesions and those from metastatic lesions (P = .63) (Table 3).

Comparison of EGFR Mutation Status Between Archival Smear Slides and Cell Blocks

To validate the mutation results obtained from archival smear slides and corresponding cell blocks, EGFR mutations in histologic specimens and in the corresponding archival smear slides and cell blocks were compared in all 43 cases for which both sample types were available. The EGFR mutation concordance rate between the histologic specimens and the corresponding archival smear slides and cell blocks was 95.3% (41 of 43 specimens) and 81.4% (35 of 43 specimens), respectively. The concordance rate was significantly higher in the corresponding archival smear slides (95.3%; 41 of 43 slides) than in the cell blocks (81.4%; 35 of 43 cell blocks; P = .04) (Table 3). The sensitivity, specificity, and accuracy were 90.4%, 100%, and 91.6%, respectively, in archival smear slides and 78.9%, 100%, and 81.4%, respectively, in cell blocks. In specimens that had a sufficient number of tumor cells, both archival smear slides and cell blocks exhibited a 100% concordance rate with the histologic specimens.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

In this study, we demonstrated a strikingly high concordance rate (91.6%) in the EGFR mutation status between histologic specimens and corresponding cytologic specimens of lung adenocarcinomas. The sensitivity, specificity, and accuracy were 90.4%, 100%, and 91.6%, respectively, in archival smear slides and 78.9%, 100%, and 81.4%, respectively, in cell blocks. All of the discordant cases contained lower DNA concentrations, tumor cell numbers, and tumor percentages. The quantity and quality of the DNA also were important, and the reduced numbers of tumor cells present in the cytologic samples can hamper the ability to obtain sufficient material for molecular analysis. A previous study evaluated the minimal number of cells required for a successful DNA sequencing analysis of EGFR, and the best results were obtained when 100 cells were used, followed by 50 cells and 30 cells.[19] In our study, mutations could not be detected in samples that contained <20 tumor cells; and, in specimens that contained >30 tumor cells, the cytologic specimens exhibited 100% concordance with the corresponding histologic specimens. The lowest tumor percentage that allowed for the detection of a mutation was 15%. This result is significantly lower than the results from previous studies, which reported the detection of mutations in cytologic samples with tumor percentages from ≥25% to 50% using a direct sequencing method.[20]

On the basis of these results, our study demonstrates that the minimum requirements for cytologic samples that allow for the successful pyrosequencing analysis of EGFR are a DNA concentration >25 ng/μL, the presence of >30 tumor cells, or a tumor percentage >30%. We recommend using the pyrosequencing method to assess EGFR mutation status in cytologic specimens that satisfy at least 1 of these 3 requirements.

The concordance rate of EGFR mutations between histologic specimens and the corresponding cytologic specimens from primary tumors was slightly higher than that from metastatic lesions. In the current study, the majority of cytologic specimens from primary tumors were obtained by FNA (31 of 36 specimens), and the majority of cytologic specimens from metastatic lesions were pleural fluids (16 of 24 specimens). The FNA specimens contained higher DNA concentrations and more tumor cells than the pleural fluid specimens (P = .019 and P = .001, respectively) This difference may be attributed to the following factors: 1) FNA targets the tumor lesions more directly—thus, the specimens may contain more tumor cell components; and 2) pleural fluid specimens contain more nontumor cell components, such as inflammatory and mesothelial cells.

Recently, the intratumor and intertumor heterogeneity of EGFR mutation within individuals has been reported.[11, 23, 24] A recent study[11] demonstrated that heterogeneity in the distribution of EGFR mutations is extremely rare in lung adenocarcinomas and that pseudoheterogeneity occurs because of a combination of mutant allele-specific imbalance and heterogeneously distributed EGFR amplification, which may give rise to the different molecular profiles observed by different studies. Therefore, we speculate that the discordant findings in metastatic lesions may be caused not by heterogeneity in the distribution of EGFR mutations but by the low quality and quantity of the materials. Further study will be required to confirm our hypothesis.

We also compared the feasibility of performing EGFR mutation analysis on cytologic specimens that were processed in 2 different ways (archival smear slides and formalin-fixed, paraffin-embedded cell blocks). The archival smear slides exhibited significantly higher concordance rates with the histologic specimens (P = .04) than with the cell blocks. However, in specimens that had a sufficient number of tumor cells, both archival smear slides and cell blocks exhibited a 100% concordance rate compared with histologic specimens. The archival smear slides contained significantly higher DNA concentrations, cell numbers, and tumor percentages than the cell blocks (P < .001). This difference may be explained by the following 3 factors: 1) Because the diagnosis of NSCLC is no longer acceptable for the patients' management, most cell block specimens were used for the diagnostic, immunohistochemical characterization of the tumor subtype, which leads to lower residual tumor cell content in the cell blocks than in the archival slides; 2) formalin fixation can lead to nucleic acid degradation[25]; and 3) the tumor cell percentages were low possibly because of the relatively high abundance of nontumor cell components within the cell blocks. The tumor cells can be selectively scraped from the smears, which can reduce the nontumor cell contamination during the DNA extraction process. Although we performed selective dissection from the cell blocks, infiltrating stromal cells and inflammatory cells may have reduced the percentages of tumor cell content because of the stochastic distribution of malignant and nonmalignant cells within the cell blocks.

In summary, we have demonstrated that cytologic specimens are sufficient and reliable for EGFR pyrosequencing analysis. In addition, the cytologic materials from both metastatic lesions and primary tumors can be used for EGFR mutation analysis.

Acknowledgments

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

We are grateful to Jeong Ok Park, MT, for her excellent technical assistance.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES

This work was supported by a grant from the Korea Healthcare Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (A111405).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgments
  8. FUNDING SOURCES
  9. REFERENCES